基于证据理论的航空发动机早期故障诊断方法  

Diagnosis method of aeroengine early fault based on the Dempster-Shafer evidence theory

作　　者：杨建平[1,2] 黄洪钟[1] 苗强[1] 范显峰[1] 王贵宝[1] 

机构地区：[1]电子科技大学机械电子工程学院,成都610054 [2]中国人民解放军95389部队,桂林541007

出　　处：《航空动力学报》2008年第12期2327-2331,共5页Journal of Aerospace Power

基　　金：国家自然科学基金(60672165);国家高技术研究发展计划(863计划,2007AA04Z403)

摘　　要：在分析航空发动机早期故障特点和数据融合的基础上,将证据理论应用于发动机故障征兆级别的数据融合中,提出了构造基本信任指派函数的方法,从而得到了更准确的早期故障征兆.在此基础上,应用模糊聚类分析方法,依据最大隶属度原则得到发动机早期故障源.该方法考虑了发动机早期故障征兆不明确、不精确、不肯定等特点,应用模糊测度来表示征兆,更符合发动机早期故障的实际情况.Based on the analysis of the characteristic of aeroengine early fault and information fusion, a method of forming the basic belief function was proposed when evidence theory was applied to information fusion of aeroengine fault symptom level. Thus, more precise early fault symptom could be attained. Moreover, the acquired fault symptom was applied to fuzzy cluster analysis to obtain aeroengine fault sources. Furthermore, the proposed method used fuzzy measure to express fault symptom, and considered the characteristics of imprecision, uncertainty and ambiguity of aeroengine early fault symptom. This is more suitable for the practical situation of aeroengine early fault.

关 键 词：证据理论 航空发动机 信息融合 早期故障 模糊聚类分析 

分 类 号：V263.6[航空宇航科学与技术—航空宇航制造工程]